CHAPTER 3: THE ECONOMICS OF COMPENSATION
Overview: This chapter sets the stage for understanding compensation by examining the field of Economics from the Middle Ages to modern times.
Many things influence the compensation decisions made in organizations. The previous chapter examined the influence of our legal structure on compensation. In this and the next chapter we see that understanding compensation is enhanced through the social sciences. This chapter will focus on Economics, while Psychology and Sociology are the topics in chapter 4. Economists view compensation as a labor market determinant. Thus, economics would have to be seen as very important to our understanding of compensation. The task of this chapter then is to specify the economic factors that determine compensation, the manner in which they do so, and the relative and absolute importance of each.
Compensation decisions are made by a multitude of employers and apply to millions of employees. Employing organizations may be classified in varying ways. One classification consists of private profit-seeking enterprises, not-for-profit organizations, and government entities, a classification that seems to result in smaller differences in employment practices over time. Another classification is by organizational size like revenue or headcount, measured in several ways, all of which result in significant wage differences. Another is by industry, which results in wage and occupational differences at least as large as those by size. Still another classification is by industrial sector: manufacturing or service, with the latter representing over 80 percent of US employment but different (often lower) wage levels. There are also differences in the locations of employing organizations, sections of the country and size of the communities, all of which affect wage decisions and levels. For our purposes, the thread that runs through all of these classifications is the presence of employers with employees who demand pay if they are to continue to work. Even volunteer organizations must provide some rewards to their members.
On the other side, employees differ by occupation, skill level, age, gender, and minority or majority status and these demographic factors influence the compensation they are willing to accept. These differences affect their reward preferences, and their views about what they contribute to the organization.
Employers use compensation as a primary financial reward to attract employees, and the amount offered is a strong consideration for employees in choosing jobs. These employment exchanges are made in a multitude of labor markets that vary greatly and are not very efficient.
Economists have long been concerned with compensation theory. They have viewed employment as an exchange of labor services for payment in money or in kind. They have distinguished labor markets from other markets. They have shown how labor markets determine wages. Wage theories designed to explain the determinants of wages and wage relationships have been developed by economists for over two centuries. Wage theory has changed with changes in the economy and with changes in pertinent wage issues. A major change in wage theory came from the understanding that different kinds of wage theories were required for different types and levels of wage problems. The general wage level of the economy, for example, may require an explanation different from those required by the average wage in an enterprise, wage rates for particular jobs, and pay structures or relationships.
This chapter first provides an historical perspective of how compensation has been seen from the perspective of today's wage issues. Then, the economic ideas about the labor market are reviewed. Finally, some of the work of institutional economists will be covered along with a discussion of unions.
Just Price Theory
Although most historical theories of wages are economic, the oldest is essentially sociological. The just-price theory, adhered to in the Middle Ages, involved setting wages in accordance with the established status distribution. Wages were systematically regulated to keep each class in its customary, and hence "right," place in society. Higher-status persons got higher wages. This emphasis on tying of wages to status, and the preservation of customary relationships, although developed in pre-industrial times, has a modern ring. These ideas are still common today. Employees and organizations are concerned with using correct comparisons to assign "proper" relationships and maintain some hierarchical order of wages in the organization. An example would be organizations' pay policies that call for supervisors to be paid more than subordinates. Equity theory (see Chapter 4) also seems based in part on such thinking. As we will discover in our discussion of wage relationships (structures), customary relationships continue to exert a powerful force.
Classical Wage Theory
During the Industrial Revolution, market forces became dominant and laissez-faire principles were invoked to free market forces from custom and regulation. Adam Smith set the stage for what is now called classical wage theory by providing a plausible explanation of the relation between the price of goods and the amount of labor required to secure them. Although he did not develop a wage theory, he made a number of observations pertinent to wages. His labor theory of value, which concluded that the full value of any commodity is the amount of labor it will buy, may be considered a theory of labor demand. But his observations on wage differentials speak to today's wage issues. He suggested that people choose the employment that yields the greatest net advantage. He proposed that there are five characteristics used to differentiate jobs and thus net advantage: (1) hardship, (2) difficulty of learning the job, (3) stability of employment, (4) responsibility of the job, and (5) chance for success or failure in the work. He also identified two quite different standards for comparing things of value: use value and market value. Use value refers to the value anticipated from use of the item. It varies among individuals and over time. Market value refers to the price something will bring. In a free market where demand and supply are equal, use value will equal market value.
It was Thomas R. Malthus's theory of population that provided the raw material for the first economic wage theory. Population, according to the theory, is limited by the means of subsistence: it increases geometrically whereas the means of subsistence increases arithmetically. David Ricardo translated Malthus's theory into the subsistence theory of wages. According to this theory, wages in the long run tend to equal the cost of reproducing labor, the subsistence of the laborer. This theory, often called the iron law of wages, indicated that little could be done to improve the lot of the wage earner because increasing wages leads only to increasing the number of workers beyond the means of subsistence.
The subsistence theory was an explanation of the general level of wages in terms of labor supply. Any increase in the wage rate above the subsistence level would induce an increase in the birth rate and therefore in the supply of labor. The expanded labor supply would force the wage rate back to the subsistence level. Any decrease in the wage rate below the subsistence level would result in starvation and a reduction in the labor supply. Although the market price of labor might temporarily climb above or fall below the natural price, the two would converge in the long run. In the industrial world, the theory erred in two ways: (1) improvements in technology have greatly increased the ease and methods by which subsistence can be attained, and (2) cultural forces have limited birth rates. Although Ricardo recognized the potential effects of the second factor, he believed that labor supply rather than labor demand would determine the general wage level in the long run.
Although the iron law of wages seems to have been repealed in the industrial world, it appears to still be in effect in many other parts of the world. Population growth holds back economic development in many developing countries. Famine still exists in the world. High unemployment in most of the industrial world and the effects of the Baby Boom on the American labor force suggest further that Ricardo had a point.
The short-term version of classical wage theory was the wages-fund theory. As described by John Stuart Mill, this theory explained the short-term variations in the general wage level in terms of (1) the number of available workers and (2) the size of the wages fund. The wages fund was thought to come from resources accumulated by employers from previous years and allocated by them to buy labor currently. Employers were thought to have a fixed stock of "circulating capital" for the payment of wages. Dividing the labor force (assumed to be the population) into the wages fund determined the wage.
The theory erred in assuming that a fixed fund for the payment of wages exists and that it accounts for labor demand. Most workers are paid out of current production. Employers balance labor costs against other costs in determining labor demand. Both employers and workers, however, often talk as if such funds exist and as if they determine the amount of labor services needed. They may also accept the implication of the theory that any gain to one group is a loss to others.
Residual Claimant Theory
Francis A. Walker's residual claimant theory may be thought of as an American version of the wages-fund theory. Here, the workers' demand for wages represents the residual claimant on output after rent, interest, and profit have been independently determined and deducted. Assigning wages rather than profits as the residual seems curious, but it does suggest that distribution of income is a matter of decision. It also permitted Walker to suggest that if labor increased its productivity without the use of more capital or land, its residual would increase - the germ of a productivity theory.
Marxian Wage Theory
The economic and social climate that produced classical wage theory perhaps inevitably produced Marxian wage theory. Marx accepted the subsistence and the wages-fund theories. He interpreted Smith's and Ricardo's labor subsistence theory of value as meaning that labor is the sole source of economic value. His explanation of the wage-setting process was that the entrepreneur collects the value created by labor but pays labor only the cost of subsistence. The difference is surplus value, roughly equivalent to profit. The existence of surplus value means exploitation of labor. Competition among capitalists, Marx argued, results in the accumulation of labor saving capital (jellied labor). This substitution of capital for labor results in technological unemployment and a reserve army of unemployed. Because labor is the only source of value, substituting capital for labor results in a falling rate of profits. The only solution for capitalists is to spend relatively more on capital and relatively less on labor. In this way, surplus value can be maintained by further exploitation of labor. This exploitation of labor in turn results in a class conflict which would, according to Marx, result in the demise of capitalism.
Although Marx's assumptions and predictions were faulty, his arguments are still voiced in parts of the underdeveloped world and by radical economists in the United States. Capital produces value, but it can be defined as embodied labor. Land also produces value. In the industrial world, labor's absolute and relative shares of the fruits of production have continually increased and real wages have risen. Although many of Marx's objectives have been realized without revolution, his argument that income distribution depends on social decisions as well as economic forces remains valid.
Summary of Historical Wage Theories
We have seen that historical theories of wages hold some implications for modern wage issues. But their emphasis on the general level of wages leaves many particular wage decisions without guidance. Further, it is instructive that neither the theories that emphasize labor supply nor those that emphasize labor demand were sufficient to provide a comprehensive model of the labor market. In fact, the theories emphasizing non-economic variables seem to yield more complete answers. The just-price theory, with its sociological explanation of wages, and the Marxian implications of the influence of the social system on wages may have more to say to wage decision makers than classical economic theory.
The values that flow from classical and Marxian wage theory continue to influence current thought. For example, the belief that attempts to improve the lot of the worker are self-defeating, and that unions aid their members at the expense of nonunion members, seem to be a legacy of classical wage theory. Also, the fear of "exploitation of labor" may be a bequest from Marx.
CONTEMPORARY LABOR MARKET THEORIES
As mentioned, labor economists in recent years have developed models of economic variables that focus on labor markets and have tested them empirically. The requirements of careful testing have meant that models of parts, rather than of labor-market operation as a whole, have been developed. Thus wage theory includes the quite well-verified models of labor supply, labor demand, the search process, collective bargaining, wage structures of various kinds, benefits, and the general level of money wages.1 Some of these models have been elaborately developed and carefully tested over time. Others have been less well developed and subjected to fewer empirical tests. In either case, we shall attempt to discuss them in a non-technical manner.
Markets consist of supplies and demands. This holds for labor markets, although they differ from other markets because of the relatively long duration of employment relationships. Labor markets do some things well and others poorly. They are reasonably effective in determining relative wage rates for different jobs and in raising wages and living standards with increases in productivity and national output. Labor markets also do reasonably well in determining relative wage levels for different occupations, establishments, industries, and regions. Labor markets sort millions of workers with varying skills and interests into thousands of different jobs.
What labor markets do poorly is regulate working conditions and personnel policies. They still set limits, because workers can quit if conditions become too unsatisfactory, but the limits are quite wide because it is difficult for the worker to determine these conditions before employment. Perhaps the major reason labor markets do poorly, in regulating working conditions and personnel policies, is that most workers regard changing employers as a last resort; they don't regard changing as an opportunity. Typically they try to stay with their present employer or to change that employer's behavior through a union. Also, labor markets almost always contain more people seeking jobs than jobs seeking workers. These shortcomings of labor markets are major reasons for the employer, union, and government rules that supplement the operation of market forces.
Labor economists now define labor supply as the amount of work supplied by a given population. From this definition, four dimensions of labor supply are usually noted: (1) the labor force participation rate, (2) the number of hours people are willing to work, (3) the amount of effort workers exert at work, and (4) the level of training and skill of workers. See www.bls.gov/data/ for more information.
Labor Force Participation Rate.
The quantity of the labor supply is termed the labor force. A person is included in the labor force if he or she is 16 years of age or older (14 with permission in the U.S. for certain jobs) varying by country, is able to work, and is either working or looking for work. The size of the labor force is determined by the population and the number in the labor force (termed the labor-force participation rate). This figure depends on labor-supply decisions within a household that are made only partially on economic grounds. These rates vary by age, gender, race, and educational attainment. The labor force participation rate has been shown to vary in the same direction as labor demand. This conclusion comes largely from studies of labor-force response to unemployment. There are arguments for two different responses. The first is that new workers try to get into the work force to help the economic situation in households where unemployment has occurred and the second is that the unemployed get discouraged if they cannot find a new job and simply drop out of the labor market. The added worker response is less than the discouraged-worker effect.2 The U.S. Government's Bureau of Labor Statistics (BLS) compiles data on the labor force participation rates of various demographic groups at http://stats.bls.gov/.
Population growth increases the potential labor force. The "baby boom" (roughly 1946-1964) greatly increased today's labor force but these workers are nearing retirement. The number of workers in the generations that follow are smaller, creating problems for the economy and organizations. The labor-force participation rate has also been changing by the increased participation of adult women and the lower rates of participation by people under 20 and over 65. Beginning with the early 2000s the expected changes are: fewer young entrants and a bulge in the 45 to 64 age bracket. More women, older workers, part-time employment, and moonlighting are also expected.3
Hours of Work.
Variation in hours of work offered by the labor force is explained by the theory of choice between work and leisure. This theory shows how a rational decision maker would respond to opportunities for work and leisure. Models may be constructed relating hours of work and household income under various assumptions. Such an analysis separates the substitution effect (an increase in the price of leisure, which encourages longer hours) and the income effect (the ability to buy more of everything, which encourages shorter hours). The sum of these two effects is the total effect of income changes on hours of work. This type of analysis may be applied to the effect of wage changes, the effect of income or payroll taxes, effects of competition in labor and product markets, personal preferences, productivity changes, non-work income, and overtime and shift premiums. Empirical work based on these analyses shows that the income effect is greater than the substitution effect. Wage increases reduce preferred hours, but not as much as non-labor income increases. Worker preferences are shown to have a powerful influence on working hours. Worker preferences are not uniform among individuals. Overtime premiums make many workers eager to work overtime. Shift premiums are usually too small to encourage work at unusual hours.
Working hours have been declining over time but have stayed near 40 hours per week. In the U.S the number of working hours since 1970 has risen for male workers, some regularly working more than 50 hours per week. This trend has been especially pronounced among highly educated, high-wage, salaried, and older men who are highly paid.4 Internationally, workers in the U.S. work more hours per week than all but New Zealand. Also, full time workers in the U.S. put in an average of 46.2 weeks of work per year which is high compared to most other countries.5
An economic theory of supply of effort is much less developed. The pace of work should be faster where hours are short. There should be an increase in the intensity of effort as real wages rise, because of a better diet and better medical care. In industrialized countries, the primary effect of the level of wages on effort operates through supervision and incentive-pay plans. But custom serves to set limits on work pace. In a way this whole text is about programs to influence the effort that employees put into work.
Human Capital Theory
The quality of labor supplies is explained by human capital theory. Experience, schooling, and on-the-job training are all forms of investment in human capital, as are expenditures designed to improve worker health. The costs of migration to labor markets offering better employment opportunities are also investments.
Human capital theory analyzes the effects of additional experience, education, and on-the-job training on the quality of the labor force. It also analyzes employee and employer decisions on investment in human capital. Private investment in schooling is analyzed by comparing the age-earnings profiles of individuals having differing amounts of schooling. The cost of continued education to the individual includes the cost of foregone earnings as well as direct costs. To be worthwhile, an investment in additional education must yield lifetime earnings in excess of these costs. The analysis usually includes discounting additional earnings at some rate of interest to yield the present value of that investment. Such analyses show that the more schooling, the higher the average wage within that age group. Also continued-education returns are greater at completion points than in intermediate years. They also show, not surprisingly, that these investments made later in life yield a lower return. Furthermore, a high rate of return could attract many more college entrants and eventually eliminate the return.
This differential for education has varied over time. Using a long range historical perspective there was a long term rise in the returns for education beginning before the Civil War and extending to the early 20th century. There followed a decline through WW II. The latter part was particularly steep and is attributed to government policies and an increased demand for low skilled labor during WW II. Since then there has been an upward shift, although not a consistent one, and the shift has not been particularly higher than pre-WW II standards.
Following a decline in the returns to education during the 1970's, the 1980's and 1990's showed a dramatic increase in the returns to education. The most common measure is the ratio of the wage of those with at least 16 years of education compared to the high school graduate. This increase has been substantial. Around 1980 the wage premium was 45% and in 2000 it was 80%. This has not been a consistent increase as about half of it occurred after 1987. As we move further into the 2000's this premium seems to be flattening out except for workers with more than 16 years of education who still are seeing an increasing wage ratio.6
One of the most prominent features of the U.S. economy since shortly after WW II has been the increasing disparity between low-paid and high paid workers. This wage inequality or disparity has spread to all sectors of the U.S. economy. It exists within and between states, industries and job categories. The phenomenon has been well studied by economists, but the reasons for it appear complex. While there is no one answer as to why this is happening the returns to advanced education are clearly a major reason.7
Another potential individual return to additional education is preferred placement in the queue for scarce, highly paid jobs. To the extent that employers use education as a selector, returns to education are increased.
Human capital theory may yield even more information on the qualitative dimension of labor supply when applied to on-the-job training. Training within organizations involves costs. To minimize those costs, organizations usually try to hire and keep experienced employees. On-the-job training can be usefully divided into general and specific training.8 General training provides transferable skills useful to the organization doing the training as well as to other organizations. Specific, customized training provides skills useful to the former. Employers pay for specific training by paying trainees more than they are worth during training. This cost can be recouped over the period of employment by paying the workers slightly less than they are worth after the training is completed. General training, however, must be paid for by the worker through lower wages during the training period.
Some conclusions about labor supply can now be stated. Decisions on labor-force participation, hours of work, and investment in human capital are interrelated. Individuals who expect to be in the labor force all of their lives have a stronger economic incentive to invest in education. People receiving on-the-job training may work longer hours. Workers with large investments in human capital are less likely to retire early. But all of these decisions are influenced by wages, costs of training, taxes, and other income.
Like all markets, economists picture labor markets as having interacting supply and demand curves that determine the price, in this case the wage. Supply curves of employee hours can be drawn at the level of the occupation, the industry, the area, or the economy as a whole. The horizontal axis is employee hours at all levels; the vertical axis is relative wages, except at the economy level, where it is real hourly wages. The supply curve for any particular occupation, industry, or area must be forward-sloping when measured against the wages in this employment relative to wages elsewhere. The reason is that a higher relative wage attracts labor into the activity for which it is paid. For unskilled occupations, labor supply can be almost flat, being elastic even in the short run. For occupations requiring specialized skills demanding a long period of training, the short-run supply curve can be almost vertical, reflecting inelasticity. Over time, if changes in relative wages attracted more people to training, the slope would decrease unless training places were fixed or people left the employment as fast as new ones were added. The supply of labor to an occupation will have an upward slope even in the very long run, for a number of reasons. Some occupations call upon innate talents. Others have non-pecuniary advantages and disadvantages that different people value differently. As an occupation expands, it must eventually recruit people who dislike it and demand higher earnings to enter it.
In general, both short-run and long-run supply curves for occupations requiring training slope upward to the right. Short-run supply curves are not perfectly vertical (inelastic). Long-run supply curves are not perfectly flat (elastic). A labor-supply curve for the economy as a whole in developed countries may be downward-sloping (or backward-bending). This occurs because of the tendency for hours per year to be reduced as real income rises with the labor force participation rate unchanged.
In economics, labor demand explains how much labor employers want to employ at different wage rates. Actually it is how many units of labor, not necessarily the same as number of employees. As such, it is a major component of wage theory in situations where market forces have influence.
The American labor force is employed in many thousands of specific jobs, in specific locations, with specific employee requirements. The pattern of employment is constantly changing. The labor force is constantly being reshaped to fit changes in the demand for labor. This constant change is due to changes in product demand, work methods, and the location of organizations. It is also a response to cyclical fluctuations in demand: the birth and death of organizations, the entrance and exit of labor-force members, and the search for better jobs.
Foreign competition and changes in consumer tastes have resulted in shifts in the relative importance of industries. One notable change has been the rise in the service sector, which now employs almost 80 percent of the labor force with the proportion still rising. These shifts have changed the occupational distribution. The demand for professional and technical skills has increased greatly, for example. The demand for clerical workers has also increased. Each of the occupational categories (unskilled, semiskilled, skilled, clerical, technical, professional, managerial) imply different levels or skills, each containing a wide range of jobs in terms of the skill and training required. Further, these differences in skill requirements within and between categories are increasing. There is a current trend toward demand for both employees with high skills and low skills, while the traditional blue-collar skilled worker is in lower demand.
Labor demand also varies by area. Job demand may be increasing in some areas (such as the Sunbelt) while declining in others (such as the Snowbelt). This rising demand in growing regions is met largely by migration from declining areas.
When labor demand shifts, people learn the jobs in demand and move to where these jobs are. The more the demand shifts, the more workers will have to change jobs. Historically, workers regard enforced job changes as disasters rather than opportunities. Such movement raises unemployment. Also, displaced workers may not have the qualifications to fill vacancies among the jobs in demand.
The generally accepted theory of labor demand is an application of the marginal-productivity theory of the demand for any factor of production. Because the demand for labor is derived from the demand for a product or service, it is almost always employed in combination with other productive factors. A demand schedule for labor is the relationship between a price (wage) and the quantity of labor needed. Demand alone does not determine the wage except where the supply schedule is perfectly vertical (inelastic), a very unusual situation. Demand has no effect on the wage where the supply curve is perfectly flat (elastic), but it does determine employment. In all other cases, wages and employment are jointly determined by supply and demand.
Quantity of labor demand may be expressed in working hours if it is assumed that wages represent the total costs of employment and that labor productivity is independent of the length of the workweek. Demand for labor in competitive markets is derived from a production function representing, for an unchanging technology, all possible combinations of labor and capital inputs. Selecting a particular production function, to produce maximum output at some fixed level of capital inputs, produces the short-run marginal (additional) product schedule of labor (see figure 3-1).
Figure 3-1. Short-run, marginal product schedule of labor
The downward-sloping portion of this curve shows how the marginal productivity schedule determines the quantity of labor demand in the short run. This downward slope represents the law of diminishing returns. Assuming that all units of production are sold at the same price (which they are under competitive conditions), the marginal-productivity schedule times price is the short-run demand for labor.
Deriving the long-run demand for labor involves assuming that the nature of capital services can be varied. Since it is possible to substitute capital for labor through changes in the amount and kind of capital equipment, the long-run demand curve for labor is flatter (more elastic). In the longer run, technology can be influenced by relative factor prices, so that new technologies are devised, resulting in more output for given quantities of inputs. Under competitive conditions, where product prices are assumed constant, an organization's demand for labor is the product of marginal physical product times price; this gives the value of marginal product (VMP). The demand curve for a labor market is the sum of the demand curves for the firms included in that market.
Since the demand for labor is derived from the demand for products or services, a shift in the demand for the products will produce a shift in the demand for labor in the same direction. Likewise, even under competitive conditions, an increase in costs in one industry may result in substitutions in production and, if price must be increased, substitution by consumers. Thus, the slope of the demand curve for product (elasticity) and ease of substitution by the producer affect the shape of the labor-demand curve. In fact, demand for labor will have more slope (more inelasticity), if it is harder for producers and consumers to substitute. Smaller ratios of labor cost to total cost (unless the producer can substitute more easily than the consumer) and more slope (more inelasticity) in the supply of other factors have the same effect.
A labor-demand curve for the entire economy can be constructed for competitive conditions by assuming that the aggregate demand for all the factors remains unchanged. This demand curve will still slope downward but will have less slope than curves at lower levels of aggregation.
Marginal-productivity theory assumes that the demand schedule is independent of the wage rate. It is sometimes argued that this assumption is incorrect, that a wage increase can increase the efficiency of workers or of management. The assumption that the demand schedule is independent of the wage rate probably has little merit in developed countries. The opposing argument, called shock theory, assumes that all organizations have slack and that when threatened increase their efficiency. Thus a new, higher marginal-productivity curve is created. Shock theory is most plausible when applied to a newly unionized employer.
Fixed Labor Costs.
Up to this point it has been assumed that the cost of labor to the employer consists only of the hourly wage. Dropping this assumption permits recognition that there are fixed costs of employment that vary less than proportionately with the hours an employee works. There are turnover costs, for example, including recruitment, selection, and training costs, as well as severance pay and increased unemployment-insurance costs. Other costs, such as some taxes on employment plus contributions to health and welfare plans, are unrelated to hours of work. Other benefits are not fixed costs unless they are based on hours worked.
Many turnover costs are related to employee skill level, and employers are more reluctant to lose skilled employees. Employer provision of greater stability of employment for skilled employees may be largely explained by these fixed costs. A similar distinction is often made between manual laborers and white-collar employees.
If employers keep workers in spite of a drop in demand because of their investment in them, the effect is to pay them above their current marginal product. This is done on the assumption that their marginal product will exceed their wage over a longer period of time. The fixed costs of employment also motivate employers to try to cut voluntary turnover. Thus employers find it to their advantage to tie some benefits and layoffs to seniority and to consider seniority in promotions.
Organizations that make extensive use of internal labor markets and have few ports of entry, are especially likely to emphasize promotion from within to protect their investment in training. Markets for workers with specialized training cannot be highly competitive, and employers can be expected to try to cut turnover in order to spread the fixed costs of employment over a long tenure. Doeringer and Piore (1971) describe the distinctive differences between the internal and external labor markets, or the "dual labor market" theory, and the implications for individual employment opportunities.9
The fixed costs of employment also partially explain the growth of temporary-help agencies. The employer pays an hourly rate above the prevailing rate for permanent employees. The worker gets less to cover fixed costs and agency fees, but can work the hours he or she prefers. An additional employer decision, affected by the fixed costs of employment, concerns demand for overtime hours. Overtime hours involve payment of a premium and often result in a loss of efficiency. But adding workers instead of assigning present employees overtime increases the fixed costs of employment.
Overall it appears that changes in the demand for labor are the most prominent reasons for the increased inequality of wages. During the past thirty years the demand for labor began to change in the following ways:
- Skill Based Technology. New technology creates change, particularly in the types and number of jobs needed to produce the organization's output. Recent years have seen a revolution in technology. Led by the computer and internet, almost all industries have overhauled their production processes by integrating information systems, which has transformed their organizational structures. These changes have disrupted the demand side of the labor market and made many workers obsolete.
Production processes are far more automated and productivity is much higher. The type of worker required has changed to one who has to have specialized skills and has to be more of a decision maker. These changes have also made production more efficient allowing the same or higher production with fewer workers.10
The impact of computers and internet has changed the nature of jobs at all levels of organizations. At lower levels certainly clerical work has changed dramatically; but higher-level jobs have also been impacted. On top of all this there has been a whole new class of jobs created: While some of these, like data entry operators are low level, but most of these jobs have been high-level jobs commanding high wages. While there is an unquestionable effect of computers and the internet, the production changes in past years have been far more extensive than just adding computers.
- International Trade. This is another factor whose effect is hotly debated. The argument for saying international trade has had a major effect is in two parts. The first relates to imports. As the U.S. has imported goods whose production requires low skills to produce, this has reduced the demand for low skilled workers. This reduction has created an over-supply of those workers that have the least ability to adapt to changes in the workforce leading to the reduction in real wages of the lowest paid workers.
The second part of the argument is the obverse of the first and deals with exports. The exports, in this case, are jobs. The most thought of example is U.S. companies establishing their production facilities in foreign countries. The more recent trend is outsourcing organizational functions, such as back office operations like accounts payable, accounts receivable, payroll, and collections. The interesting thing about outsourcing is the functions being affected include higher skilled ones like computer programming. The effect of these exports is to lower demand for this type of labor in the U.S. that in turn can have the effect of lowering wages.
- Growth of Services. The growth of services and the decline of manufacturing is not a new trend; it has been going on a long time as illustrated in figure 3. The past twenty years have seen an acceleration of this trend. For instance, in each of the recent economic downtrends new job growth has lagged at the end of the downturn. This is because when employees are laid off their jobs disappear; there is no coming back to that job when production increases. These workers need to develop new skills to enter the opportunities that are being created in other and often entirely new industries.
Figure 3-2. Change in Employment in Agriculture, Manufacturing and Services, 1900-2000
The switch to services can have the effect of increasing wage inequality. Service jobs most often pay less than manufacturing jobs. This is sometimes a lower skill problem and sometimes an industry problem. On the other hand, there are some services that employ mainly highly skilled workers, like professionals. These types of services are expanding greatly.
In summary, the demand for workers over the past 30 years has been consistently toward higher and higher skills that have both driven up the top of the wage distribution and lowered the bottom of the distribution.
The growth in temporary-help agencies, outsourcing, and contract employment during the 1990s should not be overlooked. It is estimated that over 20% of American workers could fall into these categories. Taken in combination with an economy that has mutated from an industrial and manufacturing employment base to a service base, one can envision an unsettling future. Since it takes 25 minutes to make a decision to downsize and 25 hours to communicate this to a service work force (and perhaps 25 days of pay for termination notices), the U.S. now faces a future in which it could go from 5% unemployment to 25% in less than a month. This was not so in the mid 1950's when manufacturing firms, with their inventories, capital, unions, transportation contracts, etc. existed. These firms are now offshore. Most service firms can be closed down by turning off the lights.
Search in Labor Markets
Labor markets may be viewed as a flow of people from unemployment to job vacancies where both are being constantly renewed. These two "pools" are related in that a change in one creates forces that change the other. The flow during the year is much larger than the net change over a year.
Analyses of labor markets explain how workers and employers search for and find each other. Because labor markets differ from other markets, the way in which they operate must be considered. Labor markets usually do not clear (demand doesn't equal supply), in part because wages do not respond fully to changes in supply and demand. This is particularly true in surplus labor conditions when wages do not fall as quickly as they once rose, leading to the comment that wages are "sticky on the downside." The quantity of labor supplied is composed of employment and unemployment. The quantity demanded consists of employed workers and job vacancies. It is useful to think of a labor market as a flow of people from unemployment to job vacancies where both pools are constantly being renewed. Although the pools are related, they are subject to different forces.
Search by Individuals.
Because of differences among workers and jobs, even within an occupation, and because employment usually involves an important long-term transaction, search is required by both parties. Furthermore, the search is required each time an employment exchange is made. Large differences in terms offered may occur, and substantial costs of the search process may be justified.
Even an experienced unemployed worker, looking for a job in his or her usual occupation, is unlikely to know which employers are hiring without searching. Employers will be offering different wages. Because accepting the first offer would be unlikely to produce the best bargain, the worker must hunt. Costs of search involve the worker's own time (the most important) and direct costs. These costs used to be extensive and overwhelming. Today, with an Internet connection, a worker can access literally tens of thousands of job postings. (See ERI Distance Learning Center Course 54: Online Recruiting.)
Workers start by having an idea of what wage they realistically can expect to obtain. This is called the reservation wage11. As more openings are found and investigated, the probability of getting a better offer decreases and the reservation wage is re-considered. Eventually, the worker decides that further search will cost more than the possible gain, so he or she will accept the best offer still open. The possibility that the job has been filled or that the employer's time limit for acceptance or rejection has been exceeded increases continuing search costs. If unemployment is very high, a worker may accept the first job offered. The search may be long and costly before any offer is received. Frictional unemployment (that generated by the search process) is thus related to labor demand.
Job seekers include not only the unemployed, but those who are dissatisfied with their job and are looking for a better one. High-level employees usually find another job before leaving their present one. In this case, the chief cost of search is the loss of leisure time. Accepting or rejecting job offers does not turn exclusively on the wage offered. Job security, commuting costs, advancement opportunities, benefits, and other less tangible factors are considered. If, however, the wage rate is the only certain information, it may be decisive. When job seekers enter the market with a minimum acceptable wage expectation (based upon the last job held or other information), the search may take longer. But this lowest acceptable limit gets revised, as lower offers are received, until an acceptable salary and benefits package is found.
Employers with vacancies are also engaged in search. If they find their vacancies are hard to fill, they improve wage offers, lower their hiring standards and/or expand their recruitment area. Employers can cut search (recruitment) costs by paying above-average wages, hoping to attract a pool of applicants. They can also attempt to keep present employees and to recruit friends of present employees. Most commonly, they use present employees to fill openings, so that only a few jobs are filled from external labor markets. Large employers hiring substantial numbers of people for jobs are less likely to change wage offers. They are most likely to widen their area of search, use more varied recruitment methods, or lower their hiring standards. Low-wage employers must search longer, spend more on recruitment, and perhaps lower hiring standards more often. High-wage employers can often choose from among a pool of applicants and spend less on recruiting. Employers who frequently use internal labor markets must look for promotion potential as well as present skills.
The Internet has substantially cut employers' search costs. For a small fee, employers can post jobs on websites such as www.monster.com and www.indeed.com to target either a national or a local labor market. Some sites, such as www.ladders.com, allow employers to target management level and professional job seekers. Many state government agencies have job posting boards on the Internet where employers can post job orders free of charge. For example, California has www.caljobs.ca.gov.
Employers also must decide how long to continue their search. If applicants do not meet the organization's hiring standards, the cost of lowering standards is compared with the costs of further recruiting and leaving vacancies unfilled. If the latter costs are too high, hiring standards are lowered and the best applicants are hired. When unemployment is low, search costs go up. If the organization finds outside recruiting too costly and decides to fill the job from within, this can lead to numerous moves within the organization. Because only the remaining lower-level job is eventually filled from outside, external search costs will be lower, but the costs of the internal moves must be considered. In truth, most employers do not have the time, energy, ability, and/or resources to conduct adequate searches. Staffing is often haphazard.
The recruitment methods used in the search process vary by skill level and in location covered. Employers broaden their search for incumbents for higher-level jobs. Applicants for these jobs also employ a broader search. Higher-skilled job seekers, especially managers and professionals, are often willing to move to other areas for higher wages or better jobs. But this movement is not enough to equalize wages in different areas.
Unemployment affects the search process, as mentioned. But the type of unemployment is also important. Demand-deficiency unemployment, which results from sluggishness in the economy, has different effects than frictional, structural, and seasonal unemployment. Frictional unemployment occurs because it takes time for employers and workers to find each other, even though they occupy the same markets and the same places. Structural unemployment involves a mismatch between worker skills and job requirements, or the existence of workers and jobs in different places. Seasonal unemployment is exactly what the term implies: that demand increases for fixed, often cyclical time periods. Unemployment may also be differentiated as voluntary and involuntary. Voluntary unemployment involves workers who have been offered a job they can fill, but who continue to search for a better one. Involuntary unemployment means that willing, qualified workers can't find jobs. Voluntary unemployment is frictional; involuntary, demand-deficient.
Structural unemployment is becoming more of a problem. Technological change increases productivity and requires fewer workers to do the work. Further, and more important for this analysis, new skills are required of the workers. Unlike demand- deficiency unemployment, when the economy improves and workers are then re-hired at their old job, in structural unemployment, those old jobs have disappeared and the worker must develop new skills in order to find a job.
Structural unemployment varies by place, industry, occupation, and type of worker. Much of it is consistent with fixed costs of employment. Unemployment rates are higher for mature women as a result of their less continuous labor-force participation. They are lower in high-skill occupations. Teenage workers have the highest unemployment rates. Some of minority unemployment reflects differences in educational attainment and occupations. Employment is also affected by the rate of mergers and acquisitions found within the economy. The overwhelming result of conglomerate acquisition performance is the creation of unemployment in the elimination of duplicate jobs.12 National, regional, state and metropolitan area unemployment data is available from the BLS at www.bls.gov/eag.
The seriousness of unemployment depends on its duration. Cyclical and structural unemployment are longer than frictional. Teenage workers have the most unemployment but also the shortest. Because one of the purposes of our unemployment compensation system is to make possible a longer search, improving these benefits raises frictional unemployment.
UNIONS AND COLLECTIVE BARGAINING
Unions attempt to regulate every kind of managerial action that directly affects the welfare of their membership or their own strength and security. The substantive provisions of union contracts may be said to encompass three main categories: job tenure/ job security; work schedules/ work speed/ production methods; and amount/ method of compensation. All three influence wage costs and decisions. Methods of hiring, promotion, and layoffs influence labor costs indirectly. Work speeds and production methods strongly influence labor costs. The decisions on compensation, wages and benefits, impact costs directly.
Unions try to influence employers hiring, placement, and retention decisions. They are interested in ensuring jobs for their members and in influencing the wages paid by reducing the labor supply. Their interest in union security, training programs, use of seniority in promotion and transfer, as well as discipline and discharge decisions are all tied to getting and keeping jobs and employment opportunities for their members. Although unions have always shown concern with demand for labor, there is evidence that their concern for job security and benefits is increasing. A not unusual quid pro quo for wage concessions in the recession of the 1980s and 1990s was an increase in job-security guarantees. Recent negotiations almost always revolve around benefits; particularly health benefits.
Unions try to influence all of the compensation decisions affecting their members. For example, they try, often successfully, to reduce wage differences among unionized companies in the same industry. This principle of the "standard rate" for workers doing comparable work stems partly from considerations of equity and partly from political pressures within the union. A further explanation is that a uniform wage level for companies operating in the same product market "takes wages out of competition."
It has been assumed that when unions emphasize wage demands they expect that employers will set employment on their labor-demand curve. Some unions, however, force employers to use more labor than they need at the union wage; this in turn forces them off the demand curve to the right. This practice, called featherbedding, is more prevalent in trade than in industry unions. Formal analysis of featherbedding rules shows that at some point in union demands no additional labor will be employed, that these rules do not increase total employment, and that they may force employers to change factor proportions.
Economic analysis of bargaining power shows that union power is based largely on the strike. Lesser sources are political action, boycotts, and control of the labor supply. Employer bargaining of power involves the ability to replace strikers, the capacity to carry on production, and financial resources. The balance of power differs greatly among bargaining situations. In recent years the value of the strike has been decreasing, with the changes in organization and economic competitiveness reducing the influence of unions.
Union pressure frequently forces management to increase efficiency and thus the organization's ability to pay. When this "shock effect" cannot occur because of unchangeable economic or company conditions, some of the less efficient plants may be forced from the industry. If union members in a particular plant become convinced that forcing wage equalization will cost them their jobs, they will usually vote to accept a lower wage and keep the plant in operation. This is more likely to occur when the plant is geographically isolated and members would have to travel some distance to find new employment. In such cases even national union officers may be willing to accept some departure from wage equality. In general, however, unions emphasize the principle of the standard rate rather than the principle of ability to pay.
Union Effects on Wages.
Research in this area has tried to determine whether unionism makes a difference in wages. In the process it has been necessary to try to measure and adjust for the factors just mentioned in order to isolate the independent effect of unionism. Most of these studies have found that unionism has a moderate effect, on the order of 12-25 percent. Measuring this effect is difficult because it is hard to get comparable groups to measure.13 There are considerable differentials by sub-groups such as minorities, women and lower level job holders as well as by occupation, type of organization and geography. In addition, union members are more resistant to a reduction in their wages in economic downturns.
An interesting question left by such research is why, given the tremendous differences among industries in favorability of economic conditions for union gains, the union-nonunion differential has been so modest. One possible answer is the politics of unionism, which forces union leaders to keep up with wage changes, but holds out little incentive for them to lead. Another is that equity and reasonable wage relationships act as conservative forces. Still another is that market forces limit the relative wage differences among industries, which would permit all of them to attract and keep a work force. In addition, these differentials may be biased since nonunion employers often raise wages when unions win increases. The effect of unions on nonunion wages is called the spillover effect. Indirect effects occur in tight labor markets, when nonunion employers must increase wages to compete with union employers for employees.14
The union effect also makes wage rates more rigid, because labor contracts are for fixed periods. Thus the union effect is greater in depressions and quite small in periods of rapid inflation. Union effects also vary by the type of union, the skill level of those organized and by differences in product markets. Unions of skilled workers facing an inelastic labor demand may win large wage increases with little reduction in employment. This is especially true when their wages represent a small part of total costs and where they bargain independently. However, these effects cannot occur if all crafts bargain together or follow a pattern that spreads to all crafts. In general, trade unions have not been as effective in raising wages as industrial unions; this is perhaps because union wage policies often call for reducing wage differentials, and political forces often result in the largest absolute increases for the least skilled.
As already noted, there has been a decline in the earnings of skilled and semiskilled relative to unskilled workers in the United States. Unionism may have aided this tendency, but not very much. The decline in the occupational differential has varied by industry. The principle of the "standard rate" suggests that unions attempt to eliminate the differential among regions and communities of varying size. But this depends on the market structure of the industry and the scope of bargaining. Where competition is limited to a locality and bargaining is conducted locally, there is no reason to expect wage equalization among localities. Thus, where unions have organized high-wage regions and communities, they may have widened geographic differentials in such industries.
Developing formal models of the collective bargaining process has been difficult. In most situations the parties are forced to deal with each other on terms both parties accept. Neither party is dealing with a set of fixed parameters: prices may be raised or employment may be cut, for example. In collective bargaining, there are always pressures from other sources, such as the public. Also, the bargaining involves many variables, only some of which may be quantified.
A formal bargaining theory applicable to wages has been developed. Its central proposition is that each party compares its probable gain from a strike with the expected costs. The present value of the expected gains over the contract period is compared with the expected costs. Unfortunately, the expected gains and costs cannot be known in advance. Also, a matter of principle or equity is sometimes involved. During a strike, a combination of rising costs and improved information brings the parties closer together and eventually produces a settlement. The cost of a strike to a union is higher when business is bad; the cost to management is higher when business is good.
Other Union Influences.
Unions also influence the employment exchange by trying to influence legislation. Union support for minimum-wage laws is well known. Unions have also been in the forefront in reducing hours of work. Less evident is their interest in restricting immigration. All of these efforts have had some effect on reducing the labor supply and thereby increasing the general level of real wages. Unions also attempt to influence the makeup of the reward package by pushing for benefits that improve the security of employees and the amenities at the workplace. The widespread existence of some benefits is probably due to union innovation.
Unions also greatly influence personnel policies and practices within organizations. With the exception of selection and training of new employees, all personnel programs for unionized agreements are jointly determined. In fact, the erosion of the "employment at will" doctrine in the 1980s and 1990s stems at least in part from the protection against arbitrary discharge that unionism provides.
There is evidence that concession bargaining in the early 1980s resulted in a change in union concerns and bargaining tactics. One such effect has been a demand for gainsharing, including profit sharing. More significant, perhaps, is increased union concern with employment levels. Where in the past unions have usually bargained wages and left decisions on employment levels to the employer, now unions granting wage concessions often demand detailed information on company operations. Further instances of union concern with employment levels include wage concessions in return for employment guarantees, union involvement in management decisions, advance notice to unions before plant closures or work transfers, strengthening of both supplementary unemployment benefits and severance-pay provisions, as well as employee rights to transfer to still-open plants.
These trade-offs appear to involve important management prerogatives. Probably the most important is the unions' insistence on their right to be informed of company decisions that affect the organization's prospects. Pricing policy, cash-flow decisions, and new investment decisions also have all been discussed in concession bargaining. Whether these are permanent changes in collective bargaining and whether they will spread to industries in which they do not now appear cannot be foreseen. They seem quite foreign to the ideology of American workers and to management tradition. But it is hard to believe these innovations will disappear without having some effects. These trends are more similar to the rights European works councils have with employers.
There has been a rapid decline of the labor movement for many years now. The cause of this is the severe shrinkage in certain traditionally union-strong industry groups, including most of manufacturing. While teachers unions, municipal unions, and retail unions have grown, these service industries belie the fact that union influence in America is on the wane. When General Electric's President wrote a letter to all 20,000 GE subcontractors in December of 1999 and suggested, recommended, and warned them that they should consider moving their plants to Mexico (and its non-union environment), it does not take any genius to forecast the role of manufacturing unions in the future. Current statistics on union membership in the United States are available at http://stats.bls.gov/news.release/union2.toc.htm.
Wage structure theory seeks to explain the many patterns or structures of wage in the economy: the differentials created by occupation, industry, geography, race and gender, and within groups. Particular differentials seldom appear in pure form. Also, the economic significance of some wage structures is greater than that of others.
Occupational wage differentials reflect differences among workers in levels and kinds of skill, as well as in work conditions. In a perfectly competitive market almost no industry differentials, except those that flow from different occupational mixes, would be expected. The forces that determine the wage level of an occupation change greatly with the period considered. In the very short run, the number of people qualified for all but the lowest-skill occupations is fixed, and if the demand for labor in an occupation rises, wages will rise. Thus the short-run supply is inelastic and demand determines the wage. If the time period is lengthened to include time for training people, the supply depends on tastes, training costs, and expected career earnings. Thus the long-run supply could be highly elastic, with the wage determined largely by the height of the supply curve and employment by demand.
The long-run wage structures of any two occupations may be differentiated by training requirements, worker tastes, non-pecuniary advantages, worker expectations about future earnings, and the rate used to discount future earnings. If there are no differences except in training requirements, the long-run supply curve of the higher-skill occupation will be perfectly elastic at a constant differential that, when discounted, just covers the cost of training. If non-pecuniary advantages are not equal and workers prefer the higher-skill occupation, the wage differential will not cover training costs. If workers prefer the lower-skill occupation, the differential will have to be more than enough to cover training costs.
Differences among workers in tastes and expectations would cause the long-run supply curve for the higher-skill occupation to slope upward (it would be more inelastic). For constant training costs, larger differentials would be needed to attract those whose tastes and expectations differed most from those who like the occupation the most.
The advantages of an occupation include the pay and other amenities that go with it, plus its prestige and the satisfaction of the work itself. But worker tastes differ. If workers are arrayed in order from those who like an occupation the most to those who like it the least, they will form an upward-sloping supply curve. When wages are measured relative to other occupations requiring the same amount of training, then if everyone dislikes an occupation there must be a positive wage differential; however, the size of the differential will depend on demand.
This compensating differential, however, may not be required if there is substantial demand-deficiency unemployment or if there are many unemployed immigrants or minority workers. These latter groups may bid down the wage for disagreeable work that requires little skill until it is no more than the wage for agreeable work requiring no more skill.
Occupational wage differentials may also be rents for scarce natural talents. In the case of most athletes, actors, and artists, such rents account for the differentials in earnings between the most successful and the average. If average earnings in these fields are above those in others requiring as much training but less ability, the difference also represents costs. If many people of modest talent enter an occupation, the result can be low pay and frequent unemployment. Such entrants might even drive average wages below those of other skilled occupations in which ability is less important.
The most important component of occupational wage differentials is the return on investment in acquiring skills. To acquire new entrants, a skilled occupation must return the private costs of training over the entrants' working life, discounted at their rate of return.
The historical tendency for percentage occupational differentials to decline is best explained by lower private costs of training. Governments have reduced such costs through compulsory school attendance laws and free and subsidized public education. If all the costs of training needed to acquire skill were paid by the public, occupational differentials would be much smaller and would be based on compensation for the non-pecuniary disadvantages of occupations and on rents for scarce abilities. Restricted immigration early in the 20th century was also a force in narrowing occupational differentials.
In the short run, skill differentials narrow during labor shortages. This results from more individuals shifting to higher-skill and presumably more pleasant jobs and from the lowering of hiring standards for skilled jobs.
The process of correcting unjustified occupational differentials works differently when they are too large than when they are too small. When the relative wage of an occupation is too high, it is almost never cut and for long periods more people train for the occupation than there are openings. When the relative wage for an occupation is too low, the most likely result is higher starting pay and faster promotions. The adjustment process for occupational differences occurs quickly in expanding occupations, and the difference between the pay for new recruits and that for longer-service employees is often compressed.
Race and Gender Differentials
Economic analysis of wage differentials between men and women and among the races seeks to explain the forces that cause them. It is well known that such differentials exist. Men earn more than women. Minorities earn less than Caucasians. These differentials arise in part from discrimination in labor markets and in part from productivity differences, some of which may be traced to past discrimination in the labor market and in education.
The most common forms of discrimination are refusing to hire women or minorities in jobs for which they are qualified, employing them at lower wages, or insisting on higher qualifications at the same wage. All of these practices by an employer are illegal under federal law. Benedick points out that qualifications required or preferred by employers may be only weakly justified in terms of business necessity.15 Also, because employers in the same market pay different wages for the same occupation, minorities and women are over represented among employees of low-paying employers. Occupational distributions suggest that discrimination is greater in higher-level occupations, but differences in educational attainment are a partial explanation. Women are over-represented in white-collar (especially clerical work) and service occupations, and underrepresented in blue-collar occupations (especially the crafts). Although women form a higher proportion of professional and technical occupations than men, the higher-paid professional occupations are heavily male.
Discrimination in labor markets may also be traced to prejudice by the employer and to an employer's belief that his or her employees or customers are prejudiced. A theory of discrimination developed by Becker asserts that a discriminator behaves as though the wage required to hire a woman or a minority member includes a positive discrimination coefficient based on prejudices.16 Such an employer is willing to sacrifice profits to indulge these beliefs. The wage differential between equally competent African-Americans and Caucasians in a competitive labor market depends upon the distribution of employers by extent of discrimination and the size of the minority group. The more discrimination in a market and the larger the supply of minority workers, the lower the relative wage.
Unions have affected the racial wage differential. Since unions tend to promote personnel policies that discourage managerial discretion in favor of objective factors such as seniority, the overall effect of unionism is to reduce racial differentials; the effect is greater where the proportion of African-American union members is smaller. Fewer women are union members, which has led to a widened male-female differential. Unions and women are discovering each other, and the effect of this will be to reduce this differential.
Consumer prejudice can operate in occupations where workers have contact with customers. Although it operates in both directions, the net effect may be to lower the relative incomes of minorities and women. This result would be expected if Caucasian males determine policy and Caucasians have higher purchasing power.
Gender and race wage differentials are also due to productivity differences. The clearest source of these is education, which may differ in both quality and quantity. Low educational attainment is an independent source of wage differentials. But until recently there was more discrimination against educated African-Americans, which lowered their return on their education and their investment in education.
African-Americans also receive less on-the-job training than Caucasians, a result of higher unemployment for African-Americans and less experience. Also, African-Americans have been underrepresented in occupations where experience has the greatest value.
Women's relative wages suffer from the same productivity forces and some additional ones. Cultural expectations toward women affect their investment in education. When they have young children they may work part-time or are out of the labor force for a time. These forces lower women's return on education, labor-force experience, and on-the-job training. Productivity-related factors, however, do not explain all of the gender differential in wages. See Chapter 24 for more information on the gender pay gap.
Economic analysis of the effect of anti-discrimination laws differ depending on the mix of law, employer familiarity and enforcement efforts. The effect of equal-pay laws is to reduce employment of women previously encouraged by pay differentials, assuming the barrier to employment is on the demand side. If equal-pay laws are better understood and enforced than equal-opportunity laws, the effect could be reduced employment of minorities and women. If the opposite is true, such employment could be increased. But such laws change attitudes. An early study of state fair-employment laws showed that both relative-wages and employment increased after their enactment.17 Research by Jonathan Leonard shows that the minority wage gap widened in the 1980's partly as the result of weakened federal enforcement of affirmative action and anti-discrimination laws and regulations.18 In 1998, the Equal Employment Opportunity Commission (EEOC), the agency responsible for enforcement of pay discrimination legislation at the federal level, received about 6,200 wage-related charges, approximately eight percent of all charges filed with the agency.19
Wage Differentials within Groups
Economic analysis has also been applied to wage differences within groups. There is always substantial variation in earnings within each group. Economic theories that attempt to explain within-group differences focus on the effect of training, the interaction of separate abilities, and the hierarchical structure of large organizations. Assuming that all people have the same ability and that there are no non-pecuniary advantages or disadvantages in any employment, without investment in human capital all workers would receive the same earnings. If we now assume that two forms of investment in human capital are possible (education beyond the minimum and general on-the-job training paid for in part by private means), those with the easiest access to capital will invest the most. People getting on-the-job training will earn less than the untrained during training. Those who have completed training will be more productive than the untrained and receive sufficient additional income over their working life to recoup their investment plus interest.
The effect of on-the-job training interacts with the variation in earnings by age. Earnings typically rise with age, level off in middle age, and are lower for older workers. The peak comes earliest in the lowest-skill occupations. Some of the dispersion in earnings below the earnings peak is due to differing on-the-job training and some is due to changes in ability due to aging. But there can still be substantial variation in earnings of people with quite similar education.
Another model of intra-group wage differentials permits differences in ability. This will further increase the spread in earnings if ability is used in part to select for scarce training slots. Those with the greatest ability will tend to get more training and to have higher earnings after training. But it is difficult to decide how much ability affects earnings on its own and how much it does so through additional investment in human capital.
Models that consider ability directly recognize that several different kinds of abilities affect earnings. Even if all abilities were equal, economic analysis translated into earnings shows that the process of choosing careers to accommodate tastes and talents, would tend toward distorted earnings distributions. Very able people would receive high earnings, which are in part a rent for specialized ability and in part a return on specialized training. In his study of the effect of earnings on career choices, Bok finds much of the recent growth in earnings inequality occurred primarily within occupations. In the field of medicine, for example, specialists can earn two- to four-times the pay of general practitioners, resulting in an increase in the preference of top medical students for specialty practice.20
There is a distinct difference between geographic cost-of-living versus salary- and wage-level differentials. The former reflects the demand and supply for goods and services; the latter reflects the demand and supply for labor. Supply and demand account for the setting of salary structures in over 99% of businesses. It is a rare national business that sets its pay based on cost of living.
Geographic cost of living is made up of a set of costs that include housing, taxes, consumables, services, transportation, and miscellaneous costs. Typically the former two, housing and taxes, take up the bulk of an individual's costs on a graduated basis. (That is, low-income workers will spend a higher proportion of their income on consumables than high-income workers. Higher-income workers will spend more on housing, especially in the U.S. where home ownership qualifies one for personal income tax reductions.) Cost of living reflects the demand and supply for goods and services. These costs can and do vary considerably across the U.S, see Figure 3 for some examples.
Figure 3 Cost of Living in Selected U.S. Cities
Base City: United States Average
Figure 3-3. Source: ERI Geographic Assessor
Geographic salary and wage differentials reflect the demand for and supply of labor within a given area. Should workers of a particular skill be in over abundance, the demand for that skill will find the market softening as to the price of that labor. Correspondingly, should certain skills be a rarity, the individuals holding those skills should find employers competing for their talents with higher wages.
Economic analysis of employee benefits explains how this portion of compensation affects labor markets. Employee benefits are defined as any part of compensation not paid currently in money to individual employees, but paid by employers on behalf of employees. Some of these benefits are required by law. Others are provided unilaterally by employers or as a result of collective bargaining. More common is a cost sharing approach with both employee and employer contributing to the benefits costs. As usually provided, benefits determine part of the consumption pattern of employees. They may even be useless to some employees. This is in contrast with cash, which permits workers to spend their income as and when they wish.
Economies of scale in providing benefits may permit an employer to cater to the tastes of particular groups of employees in order to cut the costs of recruiting and keeping employees. If all employers provide the benefit, however, this result may not accrue.
Benefits sometimes represent employer attempts to influence the consumption patterns of employees. Subsidized cafeterias, for example, may be motivated in part by the assumption that employee productivity is enhanced by a good lunch. Employers who give employees a choice in benefits take the risk that employees will not provide for themselves. For this reason, governments mandate benefits and employers develop a core of benefits that cover all employees.
Many benefits either escape income taxation or are taxed on more favorable terms than cash wages. Tax considerations do much to explain the growth in benefits over time and their positive correlation with earnings levels. In fact, the growth of benefits is explained more by tax treatment than by all other causes combined.
As noted in our discussion of human capital theory, employers who give their employees specific on-the-job training want to keep their employees to recover their investment. One way of doing so is tying benefits to length of service. Because specific training probably correlates positively with wages, attempts to keep employees help to explain the positive correlation between benefits and wage levels.
Unions have contributed to the growth of employee benefits. Unions may be better judges of what employees want than employers. Also, union leaders may get more credit from members by bargaining for new benefits than by bargaining for the equivalent in cash. Unfortunately, the most favorable benefits have gone to workers who already had the most. Employees of large, high-pay companies and members of the strongest unions get the most benefits. Employees of small employers usually have only government-mandated benefits. As the costs of benefits keeps rising in relation to wages they become a more contentious issue in collective bargaining.
THE GENERAL LEVEL OF MONEY WAGES
Economic analysis of the money-wage level of the economy rests heavily on what happens in labor markets. As such, the general wage level in a depression is treated separately from those in more prosperous periods.
In a depression, wage rigidity means that wages are not reduced. Keynes argued that a wage reduction in a depression is undesirable because it would force a reduction in worker consumption expenditures, which in turn would lead to a further reduction in aggregate demand.21 Such cuts could also generate expectations of further reductions of wages and prices. Keynes saw the downward rigidity of money wages as an acceptable feature of industrial economies. Given the downward rigidity in money wages, a decline in demand cannot lower all wages and prices proportionately. Therefore, it must lower real output and employment. If some product prices go down, a decrease in aggregate demand may even raise the real wages of those still employed.
The mechanism whereby the general level of money wages in a depression is determined assumes that the aggregate demand curve shifts to the left. If the supply curve does not change and the general wage level does not come down, employment is reduced by the amount of decreased demand. In this new position, the wage is above the supply curve and calls attention to the involuntary demand-deficiency unemployment.
Money wages are rigid downward because union contracts fix wages for the contract period and because unions resist wage cuts. But downward wage rigidity has existed in the United States much longer than strong unionism. One reason for this is the reluctance of employers to lose employees who have been given specialized training. Another is that workers strongly resent a wage cut. Thus nonunion employers fear that a wage cut will result in lowered productivity or a union organization drive. These reasons create a situation in which, during a depression, the employed and the unemployed are non-competing groups.
Cuts in real wages resulting from price rises are not resisted with the same force as cuts in money wages. They aren't personal and can't be blamed on the employer. Moreover, they don't involve changes in relative wages.
The general level of money wages rises every year in relation to economic activity. The rise responds to strong demand and other forces, including collective bargaining. The most common explanation of the increase in the level of money wages is the Phillips curve,22 which relates the percentage change of money wages (the vertical axis in figure 3-4) to unemployment (the horizontal axis). The argument underlying this formulation is that excess demand will cause money wages to rise, and the greater the amount of excess demand, the faster the rise. Employers raise money wages when they are short of labor. Of course, the response of employers to an excess of job vacancies is somewhat sluggish.
Figure 3-4. A Phillips Curve
The Phillips curve is convex as viewed from the origin. Frictional unemployment prevents it from ever reaching the vertical axis. Wages would be expected to rise very rapidly if demand were strong enough to push unemployment below normal frictional levels. At the left of the curve, duration of job search is failing faster than the quit rate is rising. At its far right the curve is flattened by the downward rigidity of money wages. This means that the average wage will rise as long as there is excess demand in any market.
When the curve is fitted to United States data it is found to be much less convex than the theoretical curve. However, such comparisons usually include the effects of a change in the unemployment rate and changes in retail prices: both improve the fit with the theoretical curve. This suggests that the actual adjustment process involves more than just excess demand. Price changes alter labor demand but may also affect wages directly. The effect of the unemployment change is for wages to rise faster when unemployment is falling.
The Phillips curve has strong implications for national economic policy on wages and prices. If productivity changes are assumed to be constant, any given wage change alters unit labor costs. If we further assume that prices are proportional to unit labor costs, then a given price change corresponds to each rate of change in wages. Thus, the vertical axis in figure 3-4 can be changed to the rate of price change.
In this form, the Phillips curve represents economic policy choices between rate of change in prices and unemployment. Choosing either high inflation or high unemployment is not attractive. Thus attempts are made to shift the curve to the left. One approach is to attempt to lower frictional and structural unemployment by improving the public employment service or by retraining or relocating the unemployed. Another is to attempt to lower the rate of wage and price increases.
The short-run impact of a given unemployment rate on the rate of wage increase will differ from the long-run effects. If workers expect a higher rate of inflation, they demand larger wage increases. The higher the wage increase corresponding to a given unemployment rate, the steeper the curve in the long run. The Phillips curve for the United States may have shifted to the right and become steeper in part because of the changed composition of the labor force.
The economic expansion of the1990's proved contradictory to the relationship between wage growth and unemployment predicted by the Phillips Curve. During this period, low rates of unemployment were accompanied by low inflation, suggesting that a more sophisticated relationship exists between these two variables. Several factors have been proposed as contributing to the current low-inflation economic environment, such as immigration, technology, globalization, and freer capital markets.23
The most recent great recession of 2007 to 2009, began in December 2007 with the national unemployment rate at 5.0 % and ended in June 2009 at 9.5 % unemployment. The U.S. economy has yet to return to its pre-recession condition.
CONTEMPORARY INDUSTRIAL RELATIONS THEORY
Industrial relations scholars, also known as the institutional approach to wage theory and analysis, use a variety of disciplines. The distinction between labor economics and industrial relations is not a hard-and-fast one. But institutionalists, by and large, distrust static analysis and attempts to limit conclusions to a limited set of variables. For example they have long complained of the myopia of traditional economic theory when it is applied to labor markets. Their observations of wage determination in the real world, with numerous sizes and types of employing organizations and unions, have convinced them that wage theory was attempting to answer the wrong questions and was neglecting important variables. Thus, they stress the range of variables, political, psychological, social, and ethical, as well as economic, which impinge on wage determination. Also, they have been more interested in learning about the behavior of employers and workers and their unions than in developing and testing theories. The theories they have developed speak to wage decisions encountered by the parties rather than to abstractions and tendencies.
This institutional approach is heavily empirical. The details of wage experience, the variability of wage relationships, the latitude available to decision makers, and the dimension added to wage determination by collective bargaining are regarded as important matters requiring explanation. The wage rate is seen as inseparable from considerations such as timing, amount, and distribution of wages, union recognition, labor efficiency, other wage bargains, lower-wage competitors, recruitment problems, company reputation, strikes, control over discharge, and job prospects. Institutionalists' distrust of marginal productivity analysis focus on what they consider to be neglect of, or abstraction from, important variables. They approach wage decisions as involving a variety of choices that depend on weights assigned to varying and perhaps conflicting considerations. The alternatives available to the decision makers are considered as important to study as the determinants of wages and wage relationships. The effects of wage changes, in their view, must also be considered, because the parties can alter wages by administration.
The institutional approach questions the usefulness of static analysis, wherein changes are studied individually while other considerations are held constant, on the grounds that the relationships among changes are the more important issues. The observation of wage-decision making suggests that (1) a variety of results are possible, (2) the system of relationships is itself changing, and (3) results may be cumulative rather than self-correcting. Moreover, analysis limited to one point in time may hide forces that change and interact with one another over time and may ignore the influence of past decisions and future expectations.
Even the definition of wages may be an issue. Conceptually, the wage agreement in collective bargaining may be defined as all of the clauses that hold cost implications. As such, it could include, besides the effect on labor costs, the present and ultimate costs of benefits, tactical considerations involving how and when wage increases are granted, continuing bargaining in the grievance procedure, and the way various agreement clauses are administered.
Also of importance in institutional analysis are the influences exerted by institutions and their leadership. Institutional forces are seen to modify, and at times possibly even to overshadow, economic forces. Unions, for example, are observed to have a wage philosophy with strong ethical overtones. They may stress certain standards of health, safety, and decency regardless of cost. Union leaders typically do not regard cost relationships as static and unchanging. They are skeptical about predicted economic effects of wage increases. Unions differ in philosophy and type of leadership. The impact of tradition on labor markets and the psychological makeup of union leaders may be as important as economic forces.
Employers are influenced by forces other than the economy. The large monopolistic employer may be just as concerned with the status of the organization as a dependable supplier, a considerate employer, and a wage leader. Management may treat wages and prices as institutional decisions rather than as reactions to economic forces. Both may be based in part on company convenience. In large organizations, organizational slack may permit wage increases to be absorbed until prices or efficiency can be increased. Observation suggests that wage rates and benefits are paid because other organizations do so or because they are settled by collective bargaining, rather than determined by cost considerations. Employers as well as unions differ in philosophy and type of leadership. For the smaller employer, however, economic necessity is the driving force.
Institutional analysis tends to show that economic forces set rather broad limits on wages, leaving a range of discretion. But the amount of discretion varies with the type of wage rate or wage relationship under analysis. For example, the internal wage structure of a company is largely under the control of the decision makers, the external wage structure is only partially under their control, and the general wage level in the economy is beyond their control. Wage uniformity, where it appears, usually results from conscious decisions rather than from market forces. One of the strongest forces is custom.
The observation that a wide latitude exists for wage decisions prompts study of the effects of wage changes based on the reasoning that anticipated and actual effects influence wage decisions. These effects depend on non-economic considerations as well as economic ones. For example, organizational goals, union-management relations, employee attitudes, employee comparisons, communications of wage decisions, and seniority policies represent a short list of non-economic considerations.
Institutional analysis has resulted in some theories that have attempted to explain the effects of a wide range of forces. Because such theories attempt to incorporate the conclusions of institutional analysis of labor markets and collective bargaining, they emphasize wage structures (relationships) within organizations. A range theory of wage differentials developed by R.A. Lester employs what he calls anticompetitive, impeditive, and competitive variables.24 The anticompetitive factors include institutional and ethical considerations used to restrict or prevent competition for jobs. The impeditive factors include historical, psychological, and institutional considerations that present obstacles to market adjustments by creating frictions and personal preferences of workers. The competitive forces are those that affect hiring and pay.
G.W. Taylor explained wage determination in all sizes and types of organizations in terms of consequences of nonagreement.25 The upper and lower limits of the negotiating area are the points at which the employer and the employee or the employee's representative consider the cost of settlement prohibitive. Although these limits are set by economic forces, they are not precise. They are variously conceived at the start of negotiations, they change at critical points during negotiations, and they are influenced by the bargaining tactics and the skills of the parties. Using only this one variable, consequences of non-agreement, Taylor explains wage determination in situations involving individual bargaining, management-administered wage determination, and wage determination under collective bargaining.
Taylor's formulation defines wages as all clauses in the contract affecting labor costs. It also recognizes that these clauses are affected by what goes on in preparation for negotiation, the bargaining, and the effectuation of the agreement.
Dunlop's theory of wage structure seems to provide some theoretical underpinnings for two common techniques of compensation administration, wage surveys and job evaluation.26 In this formulation, the structure of wage rates within a company, within an industry, and among different industries and regions is assumed to be a system. There are two elements in the system, a job cluster and a wage contour. A job cluster is a stable group of jobs within an organization linked together by technology, administrative arrangements, and social custom. Wage rates of jobs in a cluster are related more closely to one another than to jobs elsewhere in the firm. The structure of wage rates within a cluster consists of one or more key rates and a group of associated rates. The wage structure of a firm (or establishment in a multi-plant firm) is made up of a number of clusters. Rates of key jobs in each cluster are related to one another through technology, organization, and custom.
A wage contour is a stable group of companies linked together by a common product market, labor market, or custom so that they have common wage-making characteristics. The wage rates of particular occupations in a company are related more closely to the wage rates of some firms than to those of others. A contour for a particular group of occupations is defined in terms of both the product market and the labor market. A wage contour has three dimensions: occupational, industrial, and geographic. The organizations that form a contour belong to a particular product market and may either be located in one labor market or be broadly dispersed geographically. Wage contours do not have sharp boundaries. The product market of a company may place it in several wage contours. The contour is often composed of a wage leader and a group of associated firms, and operates through wage comparisons.
Wage-making forces operate on key rates in job clusters. These rates spread externally through the operations of wage contours and internally through the relation of key job rates to one another and to associated rates.
Salkever has offered a more varied list of wage structure determinants.27 Social inertia resulting from employment consciousness, rather than wage consciousness and resistance to change, is one example. Employer judgments of the relative importance of tasks and the learning curve associated with these tasks constitute another, as do differential unemployment rates of separate occupational groups. Once created, wage differentials persist because employees reinforce these judgments. A fourth determinant is the cost to workers of acquiring skills. A fifth is the immobility resulting from non-competing groups and the labor reserves associated with these groups. Insulation from market pressures in different product markets may affect differentials. Worker willingness to undergo training and employer ability to take advantage of the learning curve affect differentials in the long term.
This chapter has reviewed the contribution of economists to wage theory. Economists have been studying wages longer than anybody else. Historical theories of wages were reviewed for their contributions to wage issues. Labor economists' application of the tools and concepts of economic analysis to labor markets was presented as contemporary wage theory. This model building and testing shows how economic variables determine the relative wage rates of particular occupations, wage relationships, and the general level of money wages.
Finally, the contributions of institutional analysis (industrial relations scholars) to wage theory were reviewed. Although economists have contributed a great deal to our understanding of wages, it is apparent that non-economic factors are also operating. As Barbara Wooten concluded, "the contemporary wage and salary structure [is] the accumulated deposit laid down by a rich mixture of economic and social forces .... [These forces] act and react upon one another to produce a result which is quite inexplicable if either is left out of the reckoning."28
The economic environment has been affected recently by major changes both in the supply of and demand for labor. On the supply side, there is an increasing participation rate, particularly on the part of women. Men, on the other hand, have shown some decline in participation. This is especially true for older men. These changes, coupled with changes in the attitudes of younger workers, have created a significant transformation in the nature of today's labor force. The demand for labor is likewise changing. The move from a manufacturing to a service society makes many of the traditional middle-class skills obsolete. There is an increasing dichotomy in the demand for labor between the high skilled and the low skilled, such that movement from one to the other becomes more and more difficult. Matching these diverse trends is the job of the labor market: clearly an imperfect market, both theoretically and practically.
The balance of power between management and labor unions has changed dramatically in the past twenty to thirty years. Union power has been eroded by the poor economy in the late 1970s and early 1980s, losses in union membership, changing composition of the work force, and changing attitude by management and the government. Organizations efforts (supported by the government) to shift manufacturing jobs off-shore further erode union power. However, the impact of unions on compensation decisions, while lessened, has not disappeared. Some of the more dramatic changes in compensation, such as two-tiered pay systems, are a result of the changing balance of power in union-management negotiations. It would appear that, for a time anyway, unions are going to be more interested in security and benefits than in increasing wages greatly.
||Filer, R., Hammermesh, D., & Rees, A. The Economics of Work and Pay 6th ED., New York, N.Y. Harper Collins, 1996.
||Rees, Op Cit.
||DiCecio, R., Engemann, K., Owyang, M., & Wheeler, C. "Changing Trends in the Labor Force" Federal Reserve Bank of St. Louis Review, Vol. 90 #1 Jan./Feb. 2008, pp. 47-62.
||Kuhn, P. & Lozano, F. The Expanding Workweek? Understanding Trends in Long Work Hours Among Men 1979-2005. National Bureau of Economic Research, Cambridge, Mass. 2006.
||Mishal, L., Bernstein, J. & Allegretto, S., State of Working America, 2006/2007. Economic Policy Institute, Washington, D.C. 2007.
||Murphy, K.M. and Welsh, F. "Wage differentials in the 1990's: Is the Glass Half-Full or Half-Empty?", http://economics.uchicago.edu/dneal/murphy.publication.doc.
||Atchison, Thomas, "Salary Trends in the United States and Europe," Compensation & Benefits Review, Vol. 39, #1, January/February 2007, pp.29-39.
||Kessler, A & Lülfesmann, C. "The theory of Human Capital Revisited: On the Interaction of General and Specific Investments" The Economic Journal, Vol. 116 October,2006, 903-923.
||Peter Doeringer and Michael Piore, Internal Labor Markets and Manpower Analysis (Armonk, New York: M.E. Sharpe, Inc., 1971).
||Robert Valetta, "Rising Wage Inequality in the U.S." FRBSF Economic Letter, 97-25, Federal Reserve Bank of San Francisco, September 5, 1997.
||Gonzalez, F. & Shi, S, An Equilibrium Theory of Declining Reservation Wages and Learning, University of Toronto, Department of Economics, Working Papers, 2007
||David J. Thomsen, "The Analysis of Conglomerate Performance," (Eugene: University of Oregon Press, 1972).
||Branchflower, D. & Bryson, A., "What Effect do Unions Have on Wages, and Would Freeman and Medoff be Surprised,?" Journal of Labor Research,Volume xxv, #3 Summer 2004.
||Farber,"Nonunion Wage Rates and the Threat of Unionization" Industrial and Labor Relations Review, vol. 58, # 3, pp. 335-52, April 2005.
||Benedick, M., "Adding Testing to the Nation's Portfolio of Information on Employment Discrimination", A National Report Card on Discrimination in America: The Role of Testing, the proceedings of the Urban Institute's Conference on Discrimination (1999).
||J. Becker, The Economics of Discrimination (Chicago: University of Chicago Press, 1957).
||W. M. Landes, "The Economics of Fair Employment Laws," Journal of Political Economy, 76 (July/August 1968), pp. 507-52.
||Leonard, J., "Wage disparities and affirmative action in the 1980's", American Economic Review, 86 (no. 2, 1996): 285-90.
||Fay Hansen, "The Importance of Compliance," Compensation and Benefits Review, 32 (May/June 2000), pp. 16-25.
||Derek Bok, The Cost of Talent: How executives and professionals are paid and how it affects America (New York: The Free Press, 1993).
||J. M. Keynes, The General Theory of Employment, Interest, and Money (New York: Harcourt, Brace, 1936).
||A. W. Phillips, "The Relation Between Unemployment and the Rate of Change of Money Wage Rates in the United Kingdom 1861-1957," Economica, November 1959, pp. 283-99.
||Harvey Rosenblum, "The 1990s Inflation Puzzle: A Commentary," Southwest Economy, 3 (May/June 2000), www.dallasfed.org/research/swe/2000/swe0003.html.
||R. A. Lester, "A Range Theory of Wage Differentials," Industrial and Labor Relations Review, (July 1952), pp. 483-500.
||G. W. Taylor, "Wage Determination Processes," in New Concepts in Wage Determination, ed. Taylor and Pierson, pp. 83-113.
||J. T. Dunlop, "The Task of Contemporary Wage Theory," in New Concepts in Wage Determination, ed. Taylor and Pierson, pp. 117-39.
||L. R. Salkever, Wage Structure Theory (New York: Humanities Press, 1964).
||B. Wooten, The Social Foundations of Wage Policy (London: Allen & Unwin, 1955), p. 252.
Internet Based Compensation Administration
Thomas J. Atchison
David W. Belcher
David J. Thomsen
ERI Economic Research Institute
Copyright © 2000 - 2016
Library of Congress Cataloging-in-Publication Data
HF5549.5.C67B45 1987 658.3'2 86-25494 ISBN 0-13-154790-9
Previously published under the title of Wage and Salary Administration.
The framework for this text was originally copyrighted in 1987, 1974, 1962, and 1955 by Prentice-Hall, Inc. All rights were acquired by ERI in 2000 via reverted rights from the Belcher Scholarship Foundation and Thomas Atchison.
All rights reserved. No part of this text may be reproduced for sale, in any form or by any means, without permission in writing from ERI Economic Research Institute. Students may download and print chapters, graphs, and case studies from this text via an Internet browser for their personal use.
Printed in the United States of America
10 9 8 7 6 5 4 3 2 1
ISBN 0-13-154790-9 01